Understanding The Miller Analogies Test for Graduate School Admission

Measuring graduate school success can be very difficult. Every year many students apply for graduate school but only a few of those applicants are admitted. Some graduate school programs require for students to take a standardized test such as the Miller Analogies Test (MAT) as part of the admissions process. MAT scores are used to predict graduate school success. A high score on the MAT means that the graduate school applicant will probably do well in the program.

The Miller Analogies Test (MAT) was developed by Dr. W. S. Miller in 1926 (Miller Analogies Test History, 2016). The test grew in popularity in the 1940’s after the University of Minnesota began to administer it to incoming graduate students. By 1947, The Psychological Corporation began to use the MAT, and it quickly spread to other graduate programs nationwide. The MAT assesses the analytical thinking ability of graduate school candidates — an ability that is critical for success in both graduate school and professional life (MAT,1996). In theory, the MAT is supposed to identify potential candidates for graduate school whose knowledge goes beyond retaining and repeating what they learn.

Since 1947, the MAT has had many changes in its maintainers. In 1970 the Psychological Corporation was bought out by Harcourt Assessment, Inc., formerly known as Harcourt Brace Jovanovich. Harcourt Assessment, Inc. continued to administer the exam until 2008 when Pearson Assessments acquired it. Pearson Assessments is the current maintainer of the MAT. Pearson assessments has various certified testing centers across the United States where students can sign up to take the test on available dates. The cost of the test can range from and up depending on the testing center.

The test also switched from paper and pencil to computer based. One advantage of this is that upon completion, the student receives a Preliminary Score Report if they chose to. However, this score is not official because it needs to be verified by Pearson before the student receives an Official Score Repot with their MAT score. The total number of questions on the test was also changed. The test was originally 100 analogies but increased to 120. Although there are 120 analogies, scoring is still based on 100 questions, and the other 20 are experimental questions that may be used on future tests (Miller Analogies Test History, 2016). Scoring for the test has also changed. In the past students were given a raw score but that is no longer the case. Students are now given a score that is normed with a baseline group who took the test in previous years (Miller Analogies Test History, 2016). This new method of scoring allows for Pearson to adjust for difficulty variations in the test because there are multiple versions of it.

The MAT tests for analytical ability to solve problems stated as analogies. It also tests a student’s ability to recognize relationships between ideas. To do this, a certain level of higher order thinking is needed. Students who do well on the test possess this level of thinking which is an indicator for intelligence, learning and educational success. Even though the MAT is an indicator for intelligence, it is not ideal to equate a high MAT test score with an IQ score. However, a high MAT score can be an indicator of intelligence. There are some high IQ societies, such as Mensa who accept the score.

The test is administered at a Pearson Certified Testing Center by a test administrator. It consists of 120 partial analogies and there are several versions that can be administered. The structure of the test questions is as follows: “A is to B as C is to D”. The analogies cover content knowledge in six major areas: general, humanities, language, mathematics, natural sciences, and social studies. Students have 60 minutes to complete all 120 questions. The MAT is also only available in the English language. The exam has no passing score because each school has its own standards. Once scores are verified, examinees receive an Official Score Report with their MAT scaled score and percentile ranks, and a list of recipient schools indicated at the time the test is taken (MAT,1996).

Psychometric Properties

Statistical significance in testing is important because it provides evidence of reliability and validity. Thompson (2002) suggests that it is evident that statistics and formulas are a much clearer form of measurement, and that qualitative data seems to be less superior than quantitative data. Messick (1980) stresses the importance of construct validity for test use because it provides a rational foundation for predictiveness and relevance, and the importance of considering the value implications of test interpretations. With the MAT we will examine how it has managed to maintain its validity and reliability over the years as a graduate school admissions test.

The MAT is a quantitative instrument because there are no open-ended questions. All 120 questions are multiple choice and are presented as partial analogies. The questions are presented as follows: “A is to B as C is to D”. For example, Plane: Air:: Car: _______ a. Motorcycle, b. Engine, c. Land, d. Atmosphere (Sheperis, Drummond, & Jones, 2020). There is little room for confusion because each question only has one correct answer. Answers to the items on the MAT are scored electronically and stored on computer files from which Official Score Reports and Official Transcripts are generated. It is important to note that Pearson does not establish a passing scaled score or percentile rank. Pearson converts raw scores to scores on a common scale called scaled scores. The MAT scaled scores currently range from 200 to 600 with a mean (average) of approximately 400 (Pearson 2017).

There are several different types of validity evidence. Validity is important because it provides evidence on whether a test measures what it is intended to measure. In this section construct and predictive validity of the MAT will be examined. The construct validity of the MAT depends on the value of the analogy format in assessing abilities that are essential to success in graduate school (Pearson, 2017). Research from Kuncel et al. (2004) provides evidence that MAT scores are strongly related with verbal ability and had a positive correlation with the time that a person took to finish a graduate degree. Furthermore, many studies of human intelligence and reasoning have found evidence that verbal, quantitative, and figural analogies are among the best to measure for verbal comprehension and analytical intelligence (Gentner, Holyoak, & Kokinov, 2001; Gentner & Markham, 1997; Holyoak & Thagard, 1996; Sternberg, 1977). These studies suggest that there is a relationship between general cognitive ability and academic performance, and that the MAT can be a useful predictor for it.

To provide evidence for predictive validity the publisher of the MAT has conducted studies involving MAT scores and subsequent GPA. The table below shoes the correlation finding for both the 1992 Psychological Corporation and 2008 Pearson studies (* = significant at 0.01 level).

Correlation findings from these two studies suggest the consistency of both MAT scores and previously earned grade point averages in predicting subsequent performance measured by first-year grade point averages (Pearson, 2017).

Reliability of a test is just as important as its validity. It is crucial for tests to yield consistent results for them to be reliable. Reliability for the MAT is measured by its internal consistency using the KR 20 formula. Reliability coefficients range from 0.00 to 1.00. The table below represents the internal consistency of all MAT test forms administered during the 2012–2015 normative sample period for the total normative sample (Total Group) and for the six intended graduate major groups that are reported on Official Transcripts (Pearson, 2017). Reliability coefficients for the MAT fall in the high .80’s and 90’s, meaning that the test is consistent.

Cultural Perspectives

The MAT is a norm-referenced standardized test. It measures for analytical skills and academic content that is essential to success in graduate school. The MAT has several different forms that can be administer to students and it is scored electronically. Once students take the test their score is compared and ranked to test scores from other students. This is a useful test because it assesses prerequisite knowledge and skills of candidates for admission to graduate programs and it is a moderate to strong predictor of subsequent performance (Pearson, 2017). This test is most appropriate for the academic setting, specifically higher education. It is typically held by universities and other Pearson certified testing centers across the United States. The test is administered electronically, and a trained test administrator is on site to give directions for students to follow.

According to Lundberg and Kirk (2004) fairness in testing is to assess students in a standardized manner, using the same methods, content, administration, scoring, and interpretation for everyone, and that is exactly what the MAT does. However, it is difficult to determine if a student’s race influences their score. Ogbu (1981), states that Black Americans perform much lower than whites on those parts of the IQ test that require abstract reasoning and other cognitive skills associated with the higher academic performance. The MAT assesses analytical thinking which is a higher-level skill, but there is not enough research or evidence to conclude if race does have an impact of scores. Further research is needed and no explanation of persons of diversity can be concluded. After reviewing Smart and Smart (1997) it appears that the concept of DSM-IV diagnosis does not apply to the MAT.

The MAT is most generalizable to potential candidates wanting to attend graduate school. Since the test is only available in English only students proficient in the English language can take it. Age seems to play a factor in scores. A study found that younger applicants do significantly better on the test than older applicants (Thumin & Boernke, 1966). However, this is not enough information to say that the test is most generalizable to younger applicants. The MAT is least generalizable to people who are not proficient in the English language or familiar with the academic setting.

Counseling Implications

After revieing the flowing articles by Lundberg and Kirk (2004); Vacc and Juhnk (1997); Maeder (n.d.) there is no information that can be applied to the MAT because it does not have counseling implications. There are also no administrative requirements that include special considerations for diverse groups.

The cost of the MAT can vary from and up depending on the testing site.

Testing sites are certified through Pearson and test administrators do not need any special credentials other than training provided from Pearson to administer the test. The test administrator needs to have basic computer knowledge. Since the MAT is reliable and valid it is used frequently by potential graduate school candidates and graduate school programs across the United States.

My student perspective of the MAT usefulness to counselors cannot be provided with the MAT, but I can provide my personal view on its usefulness in the academic setting. I have mixed feelings about the MAT. Although it seems to be a good predictor for graduate school success, I do not believe it is a test that is made for all students. I believe that there is more than one way to measure for intelligence, and the MAT limits that. Even though I have mixed feelings, I think it is a good test to switch a student’s thinking style because it makes them think analytically.

Conclusion

The Miller Analogies Test is a norm-referenced standardized test that assesses the analytical thinking ability of graduate school candidates. This is an ability that is critical for success in both graduate school and professional life (Miller Analogies Test): 1996). The MAT is a 120 multiple choice question test made up of analogies, and students have 60 minutes to complete it. It is widely accepted in the admissions process for several graduate school programs in the United States. Its accessibility, reliability, and validity make it an ideal test for candidates.